ME/CFS Follow Up Microbiome Samples

This person has been using microbiome prescription to reduce the symptoms with success and with objective measurements of improved microbiome. His MD is willing to prescribe antibiotics and the top three items (from hundreds possible) are all used by ME/CFS specialist — indicating that the model is in agreement with clinical experience of ME/CFS specialist (a.k.a. Cross-Validation).

This is a follow up to these prior posts:

Why Follow Up Posts are important

The first item is simple, does the model and suggestion appear to work. Everything is theoretically computed. The second item is that encourages people to try suggestions

Foreword – and Reminder

I am not a licensed medical professional and there are strict laws where I live about “appearing to practice medicine”.  I am safe when it is “academic models” and I keep to the language of science, especially statistics. I am not safe when the explanations have possible overtones of advising a patient instead of presenting data to be evaluated by a medical professional before implementing.

I cannot tell people what they should take or not take. I can inform people items that have better odds of improving their microbiome as a results on numeric calculations. I am a trained experienced statistician with appropriate degrees and professional memberships. All suggestions should be reviewed by your medical professional before starting.

Comparisons between Samples

First, I do not know the best way to compare samples — what I usually do is put all of the numbers side by side. Special attention needs to be paid to Lab Read Quality. A poorer read quality results in less bacteria being identified.

Lab Quality is a measure of the total number of bacteria counted. The processing of a sample may detect just 30,000 bacteria or 300,000 bacteria. This impacts the number of bacteria detected and also the accuracy of the measures.

Lab Read Quality7.
Bacteria Reported By Lab461379479383
Bacteria Over 99%ile7533
Bacteria Over 95%ile20241113
Bacteria Over 90%ile32402123
Bacteria Under 10%ile283123237189
Bacteria Under 5%ile22266143107
Bacteria Under 1%ile16194423
Rarely Seen 1%32147
Rarely Seen 5%973314
Outside Range from JasonH4477
Outside Range from Medivere15151515
Outside Range from Metagenomics6688
Outside Range from MyBioma7777
Outside Range from Nirvana/CosmosId18182323
Outside Range from XenoGene5577
Outside Lab Range (+/- 1.96SD)14968
Outside Box-Plot-Whiskers41583833
Outside Kaltoft-Moldrup211100123111
Condition Est. Over 99%ile0000
Condition Est. Over 95%ile4311
Condition Est. Over 90%ile9657
Enzymes Over 99%ile17193010
Enzymes Over 95%ile1058221968
Enzymes Over 90%ile139126296183
Enzymes Under 10%ile783369514645
Enzymes Under 5%ile542186264423
Enzymes Under 1%ile271374986
Compounds Over 99%ile33286247
Compounds Over 95%ile140127231254
Compounds Over 90%ile346307298338
Compounds Under 10%ile310227297308
Compounds Under 5%ile211111224173
Compounds Under 1%ile132476765

The next table is also very dependent of Lab Read Quality. The apparent improvement on 12/3/2021 is likely artificial because the counts are low due to low read quality.

0 – 973245151
20 – 2912131812
30 – 39410914
40 – 496893
50 – 594872
60 – 694493
70 – 79710710
80 – 897485
90 – 99141888
0 – 987295758
20 – 2914152116
30 – 3910161414
40 – 4926143
50 – 591291710
60 – 69910107
70 – 7989147
80 – 8971554
90 – 991113109

So how to interpret this wall of numbers? People can cherry-pick the numbers to say improvement or no improvement. The difference of lab read quality is a big factor because they impact the count for most of the items above. The Outside Box-Plot-Whiskers numbers show continued improvement. In short, the changes shown were less than I was hoping to see.

There is one more method of comparison — using special studies. In this case we see the average matches. Doing a little math, the expected drop of percentage due to lab quality size between 8/31/2021 and 8/11/2022 is a 10% drop. Those that exceeded 20% are color with 😊 below. Nothing became 10% worse. Note that the 😊 also agrees with comparing to 3/25/2022 (the prior sample). Other items remained unchanged. Items reported by this person are 😧 – Strong issue, 😟 – a bit of an issue

Inflammatory bowel disease4932474743.8
Small intestinal bacterial overgrowth (SIBO) 😟51294834😊40.5
Allergic Rhinitis (Hay Fever) 😧 4127454138.5
COVID19 (Long Hauler)40244329😊34.0
Irritable bowel syndrome 😟40224430😊34.0
Alcohol intolerance or Medication sensitivities43233929😊33.5
Histamine or Mast Cell issues 😧 44174324😊32.0
Post-exertional malaise 😧 36234029😊32.0
ME/CFS without IBS 😟39223630😊31.8
Poor gut motility42214024😊31.8
Brain Fog 😧 3723333331.5
Allergies And Food Sensitivity 😧 38233625😊30.5
Cold Extremities 😧 42233322😊30.0
Intolerance of Extremes of Heat and Cold 😟38163927😊30.0
ME/CFS with IBS 😟3621362730.0
Bloating 😧 37223426😊29.8
General Fatigue 😧 3420313429.8
Unrefreshed sleep3122362829.3
High Anxiety 😟3317332928.0
Easily irritated 😟32163423😊26.3
Tinnitus (ringing in ear) 😟2415292222.5
Chronic Fatigue Syndrome (CFS/ME) 😧 2521202322.3
Lab Quality7.

What is my conclusion? Most of the measures above deteriorates into noise with the exception of data from Special Studies, where we seen improvement in many measures, but not all. In one real statistical sense this makes sense: many are based on common sense and the ones showing clear improvement on statistical significance.

Going Forward

For most of my prior posts used the logical reasoning and clinical studies (which used different labs and software than the samples that I was looking at). With the special studies, we have upped our game (potentially) – the bacteria deemed significant were determined by the same lab and software of our sample, plus the study sizes was much larger than published clinical studies — hence better detection.

To build the consensus I will use the special studies, I filtered to reported issues and high percentage of matches, namely:

Remember that most of the special studies found that infrequent bacteria with a low value was what was statistically significant. This is turning the usual logic on it’s head. As I state, this is all experimental but based on studies and statistics.

The top suggestions are below


As expected, most antibiotics and prescription drugs are to be avoided. A few with positive impact includes:

In terms of generic suggestions, rifaximin (antibiotic)s is by far the top antibiotics, cited here on Health Rising: Rifaxamin – citing use by Dr. Teitelbaum, Dr. Peterson, De De Meirleir and Dr. Myhill (all ME/CFS specialists).

In short, all of the top suggested antibiotics are applicable. My personal approach would be do all three of them in a pulse manner a la Jadin, 10 days on, 20 days off and then move to the next one.

Both above and generic suggestions have proton-pump inhibitors (prescription) being the top choice for other prescription drugs.


The top probiotics list have the usual dilemma: both e.coli probiotics and lactobacillus probiotics. It’s a dilemma because they tend to be hostile to each other. My typical rotation resolution would be 2 weeks of each and then move to the next:

Probiotics to take

I should note that some are strong to be avoided (watch out for mixtures!!!)

Probiotics to Avoid

KEGG Suggestions

The KEGG suggestions top items were the bacteria found in Equilibrium and Prescription Assist, except for the top choice, Escherichia coli. A probiotic suggested by Dr. Myhill, a ME/CFS specialist in the UK. The next common conventional items are

Bottom Line

The suggestions above were done solely from special studies. The key question is are they reasonable? I would say yes based on the antibiotics suggestions — all of them have been reported to help ME/CFS patients. We also have agreement between KEGG probiotics and these suggestions.

There is a potential conceptual symmetry between the two approaches (working off extremes and using special studies that are often dealing with rare low bacteria). Bacteria influences each other in very complex ways.

The full list of suggestions is available above.

Leave a Reply

Your email address will not be published.